OpenAI's text-to-image model, DALL-E 3, has officially been released as an API, expanding its accessibility beyond its integration in ChatGPT and Bing Chat. Continuing the legacy of its predecessor, DALL-E 3 also implements robust moderation features to safeguard against potential abuse.

The API of DALL-E 3 introduces a spectrum of output formats, with the highest resolution available at 1792×1024. Creators can explore this artistic AI tool with pricing set at a starting point of just $0.04 per image. However, the current version of DALL-E 3 does have certain limitations; it does not support the editing or variation creation of existing images as DALL-E 2's API did.

Furthermore, OpenAI has programmed the DALL-E 3 API to automatically refine prompts to enhance safety and detail, which may sometimes alter the specificity of the generated images. Despite these constraints, the DALL-E 3 API marks a significant step in making advanced AI-driven creativity more widely accessible.